Machine Learning Engineer

Overview

Build the models and pipelines behind Cognite's industrial contextualization platform, turning research prototypes into production-grade, real-time applications and the infrastructure that supports them.

What you'll do

• Build, test, and maintain robust data pipelines for large-scale industrial datasets.
• Assist in deploying data science prototypes to production.
• Write high-quality, testable Python code, primarily creating RESTful/gRPC APIs for ML capabilities.
• Containerize applications (Docker) and navigate Linux to handle massive volumes of unstructured industrial data.
• Run experiments, fine-tune existing foundational models (NLP, Vision-Language), and evaluate ML libraries to enhance document parsing and entity matching.
• Write rigorous unit/integration tests and monitor deployed models for drift and performance.
• Collaborate closely with senior staff (engineers, data scientists, product managers) to deliver features on the technical roadmap.

What you'll need

• Bachelor's or Master's degree in Computer Science, Data Science, Software Engineering, or a related field.
• 3–6 years of industry experience in software engineering with a strong focus on machine learning, or vice versa.
• Strong programming skills in Python, with a proven ability to write clean, modular, and testable code.
• Solid hands-on experience with frameworks like PyTorch, TensorFlow, Hugging Face, or LangChain.
• Experience working with large datasets, writing complex SQL queries, and hands-on familiarity with data processing libraries or frameworks (e.g., PySpark, Pandas at scale, or Dask).
• Familiarity with MLOps basics, containerization (Docker), Linux command-line tools, and fundamental cloud services (AWS, Azure, or GCP).
• Solid understanding of core computer science fundamentals, data structures, and algorithms.

Nice to have

• Academic or early industry experience with Vision-Language Models (VLMs) or Computer Vision (e.g., document layout analysis).
• Hands-on familiarity with basic RAG pipelines and Vector Databases (e.g., Pinecone, Milvus, Qdrant).
• Exposure to agentic frameworks (e.g., LangGraph).
• Exposure to modern columnar data (e.g., Parquet) or data lakes (e.g., Delta Lake, Apache Iceberg).
• Understanding of efficient querying/chunking of large unstructured data for ML.
• Experience with highly unstructured data, including text and JSON extraction from messy PDFs using OCR.
• OT dataset exposure.
• Experience serving models via backend web frameworks (FastAPI/Flask).
• Comfortable troubleshooting Linux permissions, fonts, and dependencies.

Details

• Location: Bengaluru, India.